Python: Preserve deterministic workflow event order under concurrent executor completion

[davidahmann]

Problem

Equivalent workflow runs can emit executor lifecycle events in different orders when concurrent branches complete in different schedules, which makes traces harder to compare and audit.

Why now

Tracing/debugging is a core framework value. Non-deterministic event ordering increases operator friction and undermines reproducible workflow receipts.

Expected behavior

For equivalent inputs and workflow topology, emitted executor lifecycle ordering should be stable and deterministic in the runtime event stream.

Claim-to-codepath map

• python/packages/core/agent_framework/_workflows/_runner.py
• python/packages/core/agent_framework/_workflows/_events.py
• dotnet/src/Microsoft.Agents.AI.Workflows/ExecutorEvent.cs
• dotnet/src/Microsoft.Agents.AI.Workflows/SubworkflowErrorEvent.cs

Validation requirements

• Add regression coverage that exercises concurrent branch completion and asserts stable event-order semantics.
• Apply minimal runtime ordering fix to deterministic dispatch points.
• Run targeted workflow tests in Python (and .NET tests if touched).

Evidence packet

• Commit under test: $(git -C /Users/davidahmann/Projects/agent-ecosystem/agent-framework rev-parse HEAD)
• Runtime: Python/.NET workflow test harness on macOS
• Minimal repro concept: concurrent branch workflow run repeated N times yields inconsistent executor event ordering without deterministic dispatch.